Abstract: With the development of the Internet，network intrusions are all around us.The key to the intrusion detection is that how to discover the new threat in time and make the computer immunized.To solve this problem，a novel automatic immunization model，which can form the vaccine of network virus automatically and immunize computer nodes in the whole network，has been brought out.In this model，the optimal vaccine will be obtained by the way of parallel genetic algorithm.After that，by using the network distribution algorithm，in order to keep the whole network running normally，the key node of the network will be immunized at first.

摘要： 随着互联网的发展，网络风险随处可见，联网计算机迫切需要一种及时有效的方法来检测并避免网络风险。为了解决全网对网络入侵的快速检测和快速免疫问题，文章结合人工免疫思想和并行遗传算法提出了一种能够针对网络入侵，自动提取入侵疫苗，并对全网进行免疫接种的模型。模型中采用并行遗传算法提取疫苗，融合各个种群的优良个体得到最佳疫苗，通过基于关键主机的免疫信息发布算法对主机进行免疫接种，从而达到全网的免疫效果。